Data visualization

Data visualization

Information architecture

IA

Interaction design

Interaction design

Unifying the dashboard experience for scientific researchers

The image featured at the top of the about us page #1
The image featured at the top of the about us page #1
The image featured at the top of the about us page #1
The image featured at the top of the about us page #1

MY ROLE

Product designer

TIMELINE

September 2024 - March 2025

TEAM

1 product owner, 2 developers, 3 experts

CHALLENGE

Split dashboards and poor integration with results table disrupted workflows

Due to historical reasons, NCBI Virus had two separate visualization dashboards: one on the homepage and another on a dedicated page accessible only through the results table.

Researchers couldn't find tools to visualize their filtered datasets, and when they did, the filters would reset — making the visualizations useless.

View live prototype

Split dashboards and poor integration with results table disrupted workflows

Due to historical reasons, NCBI Virus had two separate visualization dashboards: one on the homepage and another on a dedicated page accessible only through the results table.

Researchers couldn't find tools to visualize their filtered datasets, and when they did, the filters would reset — making the visualizations useless.

Watch prototype demo

Split dashboards and poor integration with results table disrupted workflows

Due to historical reasons, NCBI Virus had two separate visualization dashboards: one on the homepage and another on a dedicated page accessible only through the results table.

Researchers couldn't find tools to visualize their filtered datasets, and when they did, the filters would reset — making the visualizations useless.

Watch prototype demo

RESULTS

Partial rollout

+ 5%

Geographic visualization interactions

+ 3%

Dashboard usage

+ 3%

Temporal visualization interactions

EXPECTED IMPACT

Reduced time switching between views

Increased adoption beyond taxonomy sunburst

Improved pattern discovery through interactivity

Two dashboards became one unified experience

Before

Before: dashboard #1

Before: dashboard #2

After

After: one consolidated dashboard

PROBLEM DISCOVERY

Analyzed data and user workflows, revealing dashboard fragmentation and underutilization

01

Single widget focus:

Most researchers used only the taxonomy sunburst from the first dashboard and the geographic map from the second dashboard, leaving other visualizations practically untouched.

02

Low dashboard adoption:

The second dashboard had very low usage overall, with switching back to the Results Table page being its most used feature.

Numbers show feature usage rank (1 = most used). Researchers relied on just 2 of 15 widgets.

USER RESEARCH RESULTS

Two distinct workflows revealed why dashboards were underutilized

User research identified two dashboard workflows:

  • Dashboard-First Users: Explore visualizations first → then analyze details in the results table

  • Table-First Users: Filter data in the table → visualize patterns → return to export

Beyond the issues found in analytics, both groups faced additional problems:

03

Couldn't find dashboards

04

Couldn't see data relationships

05

Lost their work when switching to another view

06

Didn't understand widgets

Homepage dashboard undiscoverable below the fold and incomplete

Second dashboard is undiscoverable with interactivity issues

DESIGN EXPLORATION

Dashboard and table on one page versus two separate pages

I evaluated two approaches for unifying the dashboard experience.

01

Combining the dashboard and results table on one page promised seamless navigation but risked performance issues due to heavy data loads.

02

Separate pages with synchronized filters offered reliability while preserving user selections.

Developer constraints on data handling led us to choose synced filters. This balanced usability with technical feasibility.

Dashboard placement options

SUCCESS METRICS

Targeting discoverability, unity, and research impact

I set clear goals to transform the dashboard.

01

Better dashboard discoverability - Researchers can easily find and access all visualization tools

02

Unified dashboard experience - All dashboards consolidated in one location

03

Increased research impact - Researchers cite our resource in their publications and use our dashboard visualizations in their work

Dashboard page prototype demo: view live prototype

Dashboard page prototype demo

DESIGN SOLUTIONS

Combined two dashboards into one, connected it to Results Table through shared filters, and enhanced charts

I transformed fragmented visualizations into a connected data exploration experience.

01

Unified dashboards and connected with the results table

  • Combined two dashboards into one

  • Synced filters across views

  • Made chart interactivity visible

02

Improved chart functionality

  • Added search to taxonomy and host charts

  • Grouped related hosts; added a taxonomy tree for navigation

  • Explained color coding and added labels and filters to time-based charts

Explore the key design decisions

LEARNINGS

A unified dashboard reveals hidden data patterns

Designing interactive data visualizations deepened my understanding of how researchers explore complex datasets and the critical role dashboards play in revealing patterns across thousands of data points.

OTHER WORK

Verifying the end-to-end workflow with an AI-assisted prototype

From design to a working prototype, built with AI